Earlier this year VMware released a new vCloud Air service for Disaster Recovery that brings business continuity protection for your vSphere on premise Virtual Machines:
- A simple solution operated from the vSphere Web client and powered by vSphere Replication
- Agile as it can scale up quickly with additional Cloud compute, storage, IPs and direct network connect capacity up to 1000 simultaneous replications to a vCloud Air Disaster Recovery enabled Virtual Data Center
- Reduces physical infrastructure Capex
- Variable term service options minimize Opex
Version 1 includes Disaster Recovery from on premise to vCloud Air, version 2 which is being rolled out in the different vCloud Air regions includes Multiple Point In Time recovery (ability to recover to one of the 24 previous replications) and Failing Back from vCloud Air to on premise.
These new capabilities as well as the possibility to testing a fail over offer new business continuity possibilities for some Enterprise Applications.
Once the vCloud Air Virtual Data Center networking setup is completed and the infrastructure Virtual Machines such as the domain controllers are replicated application Virtual Machines replications can be set up in a few minutes with a minimum Recovery Point Objective (RPO) of 15 minutes. This is technically accomplished with tracking changes and replicating only the latest deltas.
For handling the initial replication of large storage Virtual Machines it is possible to use replication seeds through vCloud Connector , the upcoming Hybrid Cloud Manager or OVF exports. Another option to work around network bandwidth is to use the Offline Data Transfer Service. The replication can be setup to quiesce any Windows based applications leveraging Microsoft Volume Shadow Services Support. On Linux file system quiescing is supported on different versions of Red Hat Enterprise Linux and Suse Linux Enterprise Server version 11. For Applications on Linux, which are not file system based, Application Level quiescing needs to be done to ensure an initial consistent seed on the Cloud environment.
The Disaster Recovery Tenant initiates the fail over. It can be a planned fail over started from the vSphere web client in which case the last on premise changes are replicated and the Virtual Machine is shut down before the fail over happens or a disaster recovery fail over initiated from the vCloud Air web portal.
There are different strategies for giving access to the failed over applications. For internal facing applications a Virtual Private Network access can be setup, for external facing application it is possible to provide access via the vCloud Air public IPs and its NAT / firewall capabilities.
Failing back is done with a reverse replication set from the cloud to the on premise data center.
Operating the replication and Disaster Recovery processes can be automated with vRealize Orchestrator plug-in for vSphere Replication workflows. This allows for example to integrate with external systems (I.E Change Management, DNS updates for public facing applications) and adjust Virtual Machine hardware settings such as advanced network mappings and CPU / Memory resizing.
Another management possibility is to offer these Disaster Recovery workflows as a Service so application owners can request new replications for the Applications Virtual Machines they manage in vRealize Automation.
vRealize Automation Replication to vCloud Air as a service
All these recent developments make applications Disaster Recovery to the cloud more convenient. Further extension of this service will broaden the Enterprise Applications being supported.
About Christophe Decanini
Christophe Decanini is a consulting architect and a member of the CTO Ambassadors group at VMware. He joined VMware in 2007 as part of the acquisition of Dunes Technologies. Based in Gland, Switzerland, Christophe is an automation and integration expert supporting VMware customers, partners, and field resources globally. Christophe has worked on several important contributions, including product features that participated in making vRealize Orchestrator becoming a successful product. He has presented orchestration solutions at conferences such as VMworld, and is a main contributor to the www.vcoteam.info blog and the official VMware Orchestrator community. Christophe reviewed and contributed to books that cover Orchestrator, including VMware vCloud Architecture Toolkit and Automating vSphere with vCenter Orchestrator (both from VMware Press), VMware vRealize Orchestrator Cookbook, and VMware vRealize Orchestrator Essentials (both from Packt Publishing). He has 20 years of experience in IT automation, was awarded the vExpert designation for several years , holds a VMware Certified Professional 5 – Data Center Virtualization (VCP5-DCV) certification and a bachelor’s degree in computer science. You can follow Christophe on Twitter @vCOTeam.